SECURITY. INFORMATION

RESEARCH MEMORANDUM

EFFECT OF LINER AIR-ENTRY HOLES, FUEL STATE, AND COMBUSTOR

SIZE ON PERFORhUNCE OF AN ANNULAR TURBOJET COh4BUSTOR

*:AT LOW PRESSURES AND HIGH AIR-FLOW RATES e1

1: 'k? By Carl T. Norgren and J . Howard Childs 4% a' Lewis Flight Propulsion Laboratory al 4 Cleveland, Ohio

I -

+.,

NATIONAL ADVISORY COMMITTEE FOR- AERONAUTICS

.. -

WASHINGTON January 13, 1953

c

NATIONAL ADVISORY C O ” l T I W 3 FCCR AEROMAVTICS

SIZE ON PEKE’ORMANCE OF AN ANNULAR TURBOJET COMBUSTOR

AT L O W PRESSURES AND HIGH AIR-FLOW RATES

By C a r l T. Norgren and J. Howard Child6

A s part of a general program t o determine design criteria f o r tur- bojet combustors, an annular combustor was developed by ut i l iz ing the design principles evolved i n previous investigations and by mking 12 design changes t o optimize the altit.lide performance of the combustor. Although the combustor was aeveloped for l iquid fuel injection, heated liquid and vapor fuels gave higher confbustion efficiencies at the severe operating cpnditions. A t conditions simulating cruise a p e d at 80,000 f e e t . i n a typical turbojet with a 5.2 pressure r a t i o engine, the heated liquid and vapor fuels gave cor@ustion efficiencies 18 and 23 percent, respectively, above the efficiency obtained with llquid

at severe operating conditions than the efficiency of any of 14 turbojet combustors previously investigated. A t rated engine speed the e f f i - ciency w a s above 97 percent with both liquid and vapor fuels at alti- tudes up t o 65,000 feet . The combustion efficiency w a s s l ight ly increased by increasing the air-flow rate per unit conibustor f rontal area t o a value 30 percent above that used In current engfne design. A further increase in the air-flow rate t o a value 69 percent above tha t of current practice resulted in markedly lower conbustion efficiencies a t the higher fuel-air ra t ios .

e

1 fuel. With a l l three fuel types the conkustion efficiency was higher

The colribustor operated with high conbustion efficiency at condi- tions where a similar but smaller combustor would not o w a t e i n a pre- vious investigation; the better performance of the combustor may therefore be due chiefly t o i ts la rger size. Combustion efficiencies obtained in previous investigations with 14 different turbojet conibus- to r s a t comparable operating conditions show a general trend of increase in efficiency with increase in cornbustor size.

- The total-pressure drop of the combustor w a s approximately twice as great as the value obtained with some production-model combustors. The combustor-outlet temperature profiles followed the pattern generally -

2 NACA RM E52J09

desired in turbojet engines. No investigation m s made of the combustor durability, carbon-forming tendencies, or other low-altitude operating problems.

INTRODUCTION

Trends toward higher flight :speeds and h i er f l i g h t altitudes for mil i tary a i rcraf t result i n a need for larger $" higher thruet) turbojet engines and engines which operate mre ef f ic ien t ly a t high altitudes. This means that colnbustor size m u s t be increased and that the combustion efficiency m u s t be increased a t the severe, low-pressure conditions encountered with reduced-throttle operation at high altitudes. The attendant improvements i n compressor performance may be expected t o make possible higher air-flow rates per unit compressor f ronta l area. If the combustor i s not t o become the engine component requiring the greatest f'ronta,l =ea, then combustors must be developed which can produce high combustion efficiencies at high air-flow rates per unit combustor f ronta l area.

Research at the NACA Lewis laboratory on designs fo r Snnular turbo- j e t combustors (references 1 t o 3) has resulted in improved al t i tude perfommce of these combustors. . Other W A research (references 4 and 5) has shown that the use of vapor fuel In l i e u of the design (liquid) fuel in a turbo jet combustor improved .the combustion effi- ciency at severe operatlng conditiom.

The research reported herein consisted of a direct-connect duct

investigation of a one-quarter segment of a 2+inch-diameter annular turbojet combustor. This research had a fourfold objective: (1) t o develop a combustor using design techniques evolved i n previous investi- gations (references I t o 3) wbich w i l l provide. high combustion efficiency at low-pressure operating conditions; (2 ) to investigate the performance of this combustor with heated liquid and vapor fuels using the conven- t i o n a l liquld-fuel injection s y s t e m ; (3) t o determine the conibustion efficiency of this combustor at the high air-flow rates per uni t f rontal area which may be produced by future improvements i n compressor design; and (4 ) t o compare the efficiency of th i s combustor w i t h similar data obtained in a previous investigation with a similar, but smaller combus- t o r (unpublished data) i n order t o show the effect of conibustor size on efficiency.

L

A n annular combustor configuration was selected f o r this investiga- t ion because this configuration best utilizes the space available for . combustion in a turbojet engine. The fuel injection system m s made similar t o those used i n a l l annular combustors previously investigated

. at t h i s laboratory, and consisted.of hollow-cone-spray pressure atomizers -

NACA RM E52J09 3

- located at the upstream end of the combustor l iner and inject ing fuel in the downstream direction. The selection of combustor size and liner

described in the following paragraphs. I design was based on the results of previous investigations in the manner

Design modifications made with two different annular turbojet com- bustors by altering the size and arrangement of the circular a i r - admission holes in the conkustor l iner resulted i n improved conibustion efficiencies and altitude operating h i t s (reference l a n d unpublished data). With both these combustors the best a l t i tude performance w a s obtained with approximately the same arrangement of circular holes in the upstream portion of the conibustor liner. In reference 2 further design modifications were made in order to improve the combustor-outlet temperature profile. Narrow longitudinal slots f o r admission of air through the upstream portion of the l iner were shown t o permit bet ter control of the outlet-temperature profile. The combustion efficiencies obtained i n reference 2 were about the sane as the efficiencies obtained with the optimum arrangement of circular holes in this same combustor (unpublished data), indicating that the longitudinal s l o t s f o r air admis- sion did not afford any iragortant gains i n cdus t ion e f f i c i ency . In reference 3 the air-admission s l o t s were ut i l ized in the design of a larger combustor. This s lo t ted amula,r combustor of reference 3 is listed as combustor G in reference 6, where i ts conibustion efficiency i s compared with that of various other turbojet combustors. These per- formance comparisons show that the s lot ted annular combustor of refer-

ating conditions of the 14 turbojet codustors of reference 6.

c

- ence 3 produces the highest confbustion efficiencies at the severe oper-

Since the earlier investigations w i t h a smaller conibustor had shown tha t s lo t s f o r air admission provide no important increase in c d u s t i o n efficiency over the values attainable w i t h circular holes, it would appear possible to obtain conibustion efficiencies compmable wi%h those of the s lot ted &z1z1u1&T ,conibustor of reference 3 through the use of the opthum arrangement of circular holes i n a combustor of the s8me size. Previous investigations (reference 7) had shown that combustor l iners having circular holes are far less s&jec t to warping than the liners having narrow longitudinal slots for pr-y air admission. Conse- quently, the combustor selected for the investigation reported herein w-as made identical in size to the s lot ted annular combustor of refer- ence 3, and the design included an arrangement of circular holes in the upstream portion of the l i ne r w h i c h was similar to the optimum hole arrangement evolved in reference 1. Rectangular slots were used i n the downstream portion of the Ilner because previous exgerience ( u n p ~ l i s h e d data) had shown these to be a suitable means for obtaining the desired

r combustor-outlet temperature prof i le .

With this basic combustor configuration, a total of 1 2 design - modifications W&S made t o optimize the performance; that is, -t;O obtain

4 - NACA RM E52J09

- the best conibination of high efficiency, low pressure drop, and desired outlet-temperature profile. The experimental investigation w i t h the f i n a l combustor (model13) included combustion efficiencies, pressure . losses, and out-let-temperature profiles. The investigation was con- ducted with l iquid, heatehliquid, and vapor fuels and with several fuel atomizers. Low-pressure operating conditions were investigated to elm- 4 ulate high-altitude flight w i t h a*-flow rates per unit combustor f rontal % area which are typical af current engine design practice, 30 percent above current practice, and 69 percent above current practice. The con- bustion efficiency data for a range of operating conditions were gener- a l ized to f a l l on a common curve, and this correlation was used t o pre- dict the conibustion efficiency to be expected with the model 13 conibustor at various flight conditione in a turbojet engine. Comparisons were made of the performance of the model 13 combustor with shilar data obtained in previous investigations with otheF"coiiibustor6 to 6how the re lat ive performance of the model 13 conibustor and to indicate the ef?hct of combustor size on performance.

N

Installation

A diagram of the conibustor instal la t ion i s ehown i n figure 1. The combustor-inlet and conibustor-outlet ducts were connected' t o the laboratory-air supply and low-pressure exhaust sptem8, respectively. Air-flow rates and combustor pressures were regulated by remote- controlled valves upstream and downstream of the conbutor. The combustor-inlet air temgerature wa8 controlled by an e lec t r ic air heater. Fuel preheat was swplied by an electric resistance-type heater.

Instrumentation

A i r flow was metered by a concentric-hole, sharp-edge or i f ice installed according t o A.S.M.E. specifications. Liquid fuel flow was metered by a calibrated rotameter; vapor fuel flow, by a calibrated sharp-edge orifice. Thermocouples and pressure tubes were located a t the combustor inlet and outlet planes as indicated in figure 1. The number, type, and location of these instruments at each plane is indi- cated in figure 2. The cmbustor-outlet thermocouples were located at centers of equal areas i n the duct. Details of construction of the thermocougles and pressure tubes axe sham Fn figure 3 and are the same as those presented in reference 3. Pressure tubes were connected t o absolute manometers; thermocouples were connected t o a recording poten- tiometer.

NACA RM E52J09 "-

- Codustors

5

CL A t o t a l of 13 codustor configurations was investigated. Each com- bustor consisted of a one-quarter segment (90') of a single-annulus corn-

c\) bustor having an outside diameter of 2+ inches, an inside " L e r of CD 1% 5 inches, and a length from f u e l atomizers t o conibustor-outlet ther- 4 0

mocouples of approximately 23 inches. The maximum co&ustor cross- sectional area was lO5 square inches (corresponding to 420 square inches for the complete conibustor). Ten simplex hollow-cone-spray pressure atomizers (corresponding t o 40 atomizers in the complete combustor) injected the fuel in the downstream direction from the upstream end of the conbustor liner. Several sets of atomizers of different capacity were used in the course of the experimerrtal investigation.

Figure 4 shows a three-quarter cut-away view of the final codustor (model 13) and figure 5, a longitudinal cross section of the combustor.. The arrangement of air-admission holes in the l iner of mode l13 combustor is shown Fn figure 6. In the present discussion the upstream one-half of the l iner w i l l be referred to as the pr-y zone and the second one- half, as the secondary zone.

The l iquid fuel used i n this investigation was MIL-F-5624A grade 3p-4. The inspection b t a for this fuel are presented in table I. The vapor fuel was commercial propane.

PROCEDURE

Conbustion efficiency and co6bustor total-pressure loss data were recorded f o r a range of fuel-air r a t i o a t tlie following conditions:

Condi- t ion

A B C D E

Conbustor-inlet t o t a l pressure

(in. Hg absolute)

15 8 5

15 15

Codustor- ~ i r - f l o w ra te inlet t o t a l per unit cam-

268 268 268 268 268

2.14 1.14 0.714 2.78 3.62

Simulated

tude in ref- erence engine at cruise r p m

f l i gh t &ti-

( f t

56,000 70,000

' 80,000 56,000 56,000

~

?Based on maximum combustor cross-sectional mea.

6 llllllc W A RM E52J09

5

These conditions sirmzlate operation of the combustor in a reference tur- bgjet engine, which is a typical 5.2 pressure ratio turbojet, operating a t a Mach nuniber of 0.6. The cruise speed of the engine is assum=d t o . be 85 percent of the rated ro t a r speed. Test conditions A through C require air-flow rates per unit .conibustor frontal. area which are typical of current turbojet engines. Test conditions D and E require air-flow rates which are 30 percent and 69 percent, respectively, above current practice.

N 4 0 u)

In the pre1imina-y research necessary t o evolve the final conibustor design, limited data were recorded with JP-4 fuel and 10.5-gallon-per- hour atomizers i n each combustor at one or more of test conditions A, B, and C . With the final (model 13) combustor, more extensive data were recorded, as indicated in the following table:

Fuel Conditions Fuel Fuel atomizer capacity Bpray

angles (*g)

JP -4 A,B,C,D,E 10.5 I 60 3.0 A,B,C 60

&Rated at 100 a / s p in. pressure differential;

%uel temperature, 3000 F. 1iqUia fuel.

Conibustion efficiency was computed as the percentage r a t i o of actual to theoretical increase in enthalpy from the conibuetor-inlet t o 'the combustor-outlet instrumentation planes using the method of reference 8. For calculation of conibustor-outlet enthalpy, the temgerature was com- puted as the arithmetic mean of the 30 outlet thermocouple indications; no corrections were made f o r radiation or velocity effects on the ther- mcouple indications.

Combustor reference veloctties were computed from t he a i r mass-flow rate, the combustor-inlet density, and the maxhum combustor cross- sectional area (105 sq in. ). The total-preseure loss was. computed as the dimensionless r a t io of the total-pressure loss t o the combustor ref- erence dynamic pressure using the naethOa of reference 3. The radial

*

I

NACA RM E52J09 - 7

- distribution of temperatures at the combustor outlet was determined at each test condition investigated and at two d u e s of combustor temper- ature rise (approximately 680° and ll8Oo F, the required values a t 85 and 100 percent rate& speed in the reference turbojet engine at alti- tudes above the tropopause). The terqperature at each of the 5 rad ia l

each radial position (see fig. 2(b)). The temperature rake at each side w a l l of the conibustor was not included in these average tenperatures in order to minimize the effects of the side w a l l s on temperature readings. The optimum conibustor-outlet radial temperature prof i le was considered to be that shown in f igure 7 ; this temperature profile represents an approximate average of those profiles required or desired in various turbojet engines.

,. 2 8 positions was computed as the average of four thermocouple readings at

The results obtained i n the investigation of the preliminary com- bustor configurations are discussed briefly in appendix A. The f o l l o w - ing results were obtained w i t h the final (mde l13) combustor. The model 13 conibustor was considered to be a near-ogtirmrm design for the particular co~.&ustor size and shape, fuel, and fuel Fnjection system

it gave the best over -a l l performance of the 13 combustor configurations investigated. The experimental data for the model 13 conibustor are pre-

appendix B.

It which w e r e selected and f o r the particular test conditions investigated;

- sented in table II. A tabulated list of all sy&ols is presented in

Conibustion Eff iclencies

The combustion efficiencies obtained with liquid, heated liquid, and vapor fuels are presented i n figures 8, 9, and 10, respectively, fo r a r a g e of fuel-air r a t i o at each of three inlet pressures ( test conditions A, B, and C ) . The data of figures 8 and 9 were obtained w i t h lO.5-gallon-per-hour, 600 fuel atomizers. These sane atomizers were used with vapor fuel to obtain a part of the data of f igure 10. Addi- t i o n a l data are shown i n figure 10 f o r 30-gaUon-per-hour, 70' atomizers.

Duplicate data were. recorded at a few values of fuel-air r a t i o at test conditions B and C, and these check data are indicated by t a i l ed syrribols in f igure 8. The check data showed an average deviation of 22 percent.

c Figure Ll presents conibustion efficiencies a t the same test condi- t ions as the preceding figures (teBt conditions A, B, and C ) with l iquid fuel and improved f u e l atomization through the use of 3. O-gallon-per- hour, 60' atomizers. -

8 NACA RM E52J09

Figures 1 2 and X5 show combustion efficiencies obtained at three air-flow rates ( t e s t conditions A, D, and E) with liquid and v q o r fuele, respectively. The data of figure 12 were obtained w i t h 10.5-gallon-per- A.

hour, 600 fue l atomizers. The data of figure 13-for vapor fue l were obtained w i t h two atomizer capacities, 30-gallon-per-hour and 60-gallon- 2 per-hour atomizers. The 60-gallon-per-hour atomizers were required t o 0 obtain the higher fuel-air ratios a t test condition E without causing 4,

the pressure drop in the-fuel InJectors t o exceed the propane fuel supply pressure, which w a s approximately 80 pounds per square inch gage.

Pressure Losses

The pressure losses through the combustor are presented i n f i g - ure 14. The dimensionless r a t i o of the total-pressure drop t o the ref- erence -c pressure m/g, is plotted as a function of the combustor-inlet t o conibustor-outlet density r a t i o pdpz . The expected straight-l ine relation is obtained, but some separation OCCUTS between data recorded at different pressures.

Outlet Temperature Profiles

Figures 15(a) and 15(b) show combustor-outlet radial temperature profiles with l iquid and vapor fuels, respectively, at each condition investigated and 4t two values of co&us%or temperature r ise (approxi- mately 680° and U8O0 F) . The desired temperature profile from figure 7 is included for comparison.

DISCUSSION

Fuel State

A comparison of the curves. from figures 8 t o 10 i s presented i n figure 16 t o show the effect of: fuel state on combustion efficiency. O n l y slight. differences exist among the liquid, heated liquid, and vapor fuels a t the two higher pressures (conditions A and B). A t the most severe t e s t condition investigated, corresponding to cruise speed at-- 80,000 feet altitude (condition C ), the' conibustion efficiency w i t h the vapor f -1 w a s 23 percent above that w i t h liquid fuel and the eff i- -

ciencg w i t h heated liquid fuel was 18 percent above that with Liquid fuel; these comparisons are made a t t h e fuel-air ratios required with each f u e l t o give the 680' F conibustor temgerature rise required for cruise speed in the reference e.ngine at 80,000 feet a l t i tude. These values of fue l -a i r ra t io axe indicated by circular symbols on each of the curves for test condition C in f igure 16.

2B NACA RM E52J09 9

t

The curves f o r l iquid and vapor f u e l at the highest air-f l o w r a t e (condition E, f igs. 1 2 and 13) show that vapor f u e l also provides a

condition. The higher co&ustion efficiency of the vapor fuel a t severe

gations of l iqu id and vawr fuels in turbojet co&ustors (references 4

I. significant increase in combustion efficiency at this severe operating

[u operating conditions is i n accord with the results of previous investi-

E: and 5). 4

A comparison of the data of figure 1l with the data of figure 8 shows that the snaaller capacity atomizers gave no significant increase i n efficiency a t any of the conditions investigated, and that they caused a slight decrease i n efficiency at the lowest pressure (condi- t ion C ) . Thus the higher combustion efficiencies obtained with heated liquid and vapor fuels were not obtained by finer mechanical atomiza- t ion of the l iquid fuel.

Performance Comparisons

Most turbojet combustor Investigations have been conducted a t test conditions simulating operation of each canibustor in its particular engine; consequently, it is d i f f icu l t t o f ind data obtained w i t h d i f - ferent conibustors at identical operating conditions. The conibustion parameter Vr/pi Ti (where Vr is the combustor reference velocity in ft/sec,. calculated from inlet density, mass-flow rate, and maximum cam-

i n lb/sq f t absolute; and Ti i s the combustor-inlet temperature i n ?R) can be used, however, t o reduce the combustion efficiencies obtained a t any test conditions to a single curve for each combustor and each fuel, as shown i n reference 6. A comparison of these curves therefore affords a comparison of the performance of various conibmtors even though the experimental conditions investigated might be different for the various combustors.

1

I bustor cross-sectional area; p i is the coDibustor-inlet static pressure

In figure 17 are plotted the cambustion efficiency data, of f ig - ures 8, 9, 10, 1 2 , and 13 as a function of the co&mtion parameter Vr/pi Ti. This f o m of the parameter is used herein rather than i ts reciprocal pi Ti/Vr as derived in reference 6 because the resulting correlation curves do not have the extreme curvature noted in refer- ence 6. The efficiency data of figure 17(a) are fo r a co&mtor tem- perature rise of 6800 F, which is the value required in the reference engine at cruise speed and altftudes above the tropopause. Fig- ures 17(b) and l7(c) present sFmi laz data for values of temperature

r i s e fo r 75- and 100-percent rated speed at al t i tudes above the tropo- pause. These values of required temperature rise were obtained from

z rise of 402O and ll8Oo F, which are the required values of temperature

-

10 NACA RM E52J09

. engine performance curves which were extrapolated to the higher altitude conditions by assuming constant efficiencies of engine components other than the combustor. Figure 18 presents a comparison of the curves of figure 17 for the model 13 combu$tor with liquid, heated liquid, and vapor fue ls with similar curves f o r two of the better combustors (com- bustors G and L) reported in reference 6. Combustor G from reference 6 is the s lo t ted annular combustor of reference 3, which produced the highest combustion efficiencies of the 14 turbojet combustors reported in reference 6. Combustor L from reference 6 is one of the better production-model combustors. Figure 18 shows that with any of the three f u e l types investigated, the modei13 combustor produced a higher com- bustion efficiency at severe operating conditions than any of the 14 combustors of reference 6.

Estimated Flight Performance

Figure B(a) shows the estimated combustion efficiency of the model 13 combustor wFth l iquid-fuel at various flight conditions in the reference turbojet engine. The curves of constant conibustion efficiency were obtained as follows : At each flight condition indicated on f ig - ure 19(a) by a circular synibol, the required temperature rise and the value of-the conibustion parameter Vr/pi T i were obtained from the engine performance curves. For each value of the conibustion parameter thus obtained the correspondfng efficiency was then obtained from f ig - ure 17 f o r values of temperature -rise above and below the required value. The efficiency at the required temperature r i s e was then obtained by interpolation. These values of combustion efficiency were next indi- cated in figure 19(a) beside the appropriate cfrcular symbols. Finally, the constant efficiency curves were drawn t o f i t the pattern indicated. by these circular symbols.

The three rectangular data points in figure 19(a) at 85 percent rated speed represent.actual experimental data where the test conditfons accurately simulated flight operation a t the conditions indicated on the figure. The combustion efficiencies listed beside each of these three data points match well w i t h the values expected from interpolation between the curves, indicating the val idis . of figure 19(a). The curve8 of f igure 19(a) show that w i t h l iquid fuel the model13 combustor opera- ted at efficiencies above 97 percent--- t o =.altitude of 65,000 f ee t at rated engine speed.

Similar data m e presented in figure 19jb) for the model 13 combus- tor operating with vapor fuel. Again the circular symbols represent data calculated by use of figure 17, and the square symbols represent .

experimental data. With vapor fue l the model 13 combustor operated a t efficiencies above 97 percent up t o an altitude of 65,000 f ee t at rated engine speed.

4 R)

0 CD

MACA FtM E52J09 ll

* With both liquid and vapor fuel the model 13 combustor supplied

sufficient temperature rise to operate the turbojet engine at 85 percent rated speed and BLL al t i tude of 80,000 feet . Thus the altitude operating limits of the engine would l i e above 80,000 f e e t a t 85 percent rated

-

Po speed.

53 4

High Air-Flow Rates

The conibustion efficiency of the model 13 conibustor was s l ight ly increased with both liquid and vapor fuels when the air-flow rate per unit frontal area was increased to a value about 30 percent above current design practice (figs. 1 2 and 13). This result is contrary to the usual trend of decreased efficiency with increased velocity which has been generally noted in previous investigations. When t h i s a i r flow was further increased to a value about 69 percent above current practice, the efficiency with both fuels w a s markedly decreased a t high fuel-air ratios. The vapor f u e l gave higher efficiencies than the liquid fuel at. the highest air-flow rate, indicating that a conibustor designed t o make bet ter use of the vapor fuel might constitute one means f o r obtaining high efficiencies a t higher air-flow rates.

n

Combustor Size

.. A development program sirmtlar t o that reported herein was pre- viously conducted with an annular combustor of smaller size (unp&lished data). The model13 confbustor evolved herein and the smaller conibustor previously investigated are similar i n many respects: (1) both confbus- tors are annular combustors w i t h similar fuel injection systems; (2) both combustors have approximately the same arrangement of circulax holes in the upstream end of the combustor l iner ; and (3) both combustors produced higher efficiencies than the various other conibustor configurations which were investigated in their respective development programs. These com- bustors differ p r h a r i l y in size. A conparison of the performance of these two combustors should therefore provide some indication of the effect of combustor size on performance.

A significant comparison of the combustion efficiency of these two combustors is not available because the combustion efficiencies of the smaller combustor were measured only at favorable conditions correspond- ing t o l o w values of the conibustion parameter Vr/pi TF where most com- bustors give efficiencies near 100 percent. The best available perfor- mance comparison between these two conibustors i s therefore shown i n figure 20. The altitude operating limits of the smaller combustor appear as a straight l ine. The experimental data obtained with the model 13 combustor are also included in figure 20 t o show that the model 13 corn- bustor operated with hi& efficiency over a wide range of conditions where the smaller conibustor would not operate.

-

12 NACA RM E52JO9

* The indicated effect of combustor size may possibly be the result

of l iquid fuel "wash" on the walls of the combustor l iner . Reference 9 shows air- and fuel-flow patterns in a turbojet combustor with no burning occurring, and an appreciable quantity of the l iqu id fue l impinges on the walls of the liner and flows along the walls as a continuous l iquid film. Liquid fue l "wash" along the walls of the l ine r has also been observed under burning conditions i n some combustors. Large quantities of l iqu id fue l on the walls would be expected t o result i n a decrease in conibustion efficiency. Since the fuel atomizers used in t he smaller com- bustor had the same flow capacity and spray angle as those used in the model 13 conibustor, it would be expected that they would came much greater quantit ies of l i qu id fue l t o impinge on the w a l l s of the smaller combustor.

Further indication of a possible important effect of c&ustor size on combustion efficiency is shown -in figure 21. Values of combustion efficiency obtained in previous investigations with 14 different turbo- jet cambustors at operating conditions .of equal severity (Vr/pi Ti = 100X10'6) are plotted in figure 2 1 as a function of a com- bustor hydraulic radius. The efficiency data for these fourteen combus- to r s were taken from the curves of reference 6. The combustor hydraulic radius i s defined as the ra t io of the cross-sectional area ineide the- - combustor l i n e r t o the wetted perimeter of the combustion zone a t the point where the undisturbed fuel .spray would touch the l iner walls. This hydraulic radius is also the voltme-to-surface rat io per unit length of the combustion zone a t the point where the fuel is dispersed across the, combustion zone. A combustor hydraulic radius w-as selected as an index of 'the effect of combustor size on efficiency because the surf'ace-to- volume r a t i o of small-scale combustion apparatus is known t o have an important e f fec t on flammability.limits, as shown in reference 10. The point where the undisturbed fuel spray would touch the l i ne r walls was arbitrari ly selected as the plane at which the hydraulic radius would be evaluated; a better correlation of the data of figure 2 1 might possibly be obtained by using values of hydraulic radius which are evaluated dif- f erently.

Figure 2 1 shows that the data points for most of the 14 combustors of reference 6 f a l l within 210 percent of a .single curve, and t h i s curve shows an increase in combustion efficiency with increase in combustor . s ize throughout the range of size investigated. The data point f o r the model 13 combustor falls near the upper end of the curve i n figure 21. Other design factors besides that of size are, of course, important, and th i s accounts fo r some of the scatter of data points in f igure 21. For example, conibustors C and D are ident ical except for changes in t he p r i - mary air holes in the l iner, and combustors I and J are identical except for changes in the fuel atomizer.

.

NACA RM E52J09 13

s Pressure Losses, Outlet Temperature Wofiles, and Other

Performance Characteristics

Y For isothermal (no conibustion) flow, the average value of m/qr T 9 in figure 14 is about 25; this value is approximately twice as great as

the corresponding value for some production-model codustors. The design changes which were made in an attempt to reduce this pressure drop were unsuccessful, as noted in appendix A; with a more extensive effort it might, however, be possible to significantly reduce this pressure drop.

1

The codmstor-outlet temperature profiles in figure 15 axe similar to the profile generally desired in turbojet engines, with low tempera- tures nees the blade hub and blade tip positions and & =bum tempera- ture at about 85 percent of the blade height. The maxFmum circumferen- tial scatter of individual thermocouple indications w-as st200° F at any radial position. The outlet-temperature profiles for the model 13 com- bustor were not so good as those of some of the preliminary conibustors investigated (for -le, see appendix A) because the model 13 conbus- tor was evolved as the result of design compromises to obtain the best over-all conibination of high efficiency, low pressure drop, and a desired pattern of outlet temperatures.

The low altitude performance of the combustor was not investigated; consequently, little is known regarding its durability or carbon- deposition characteristics. No carbon deposits or warping of the liner were observed during this Fnvestigation, but the test conditions were not in the range where these problems are severe.

SUMMARY OF RESULTS

An annular turbojet conibustor was developed to give improved per- formance at high altitudes by utilizing the design principles evolved in previous investigations to obtain the basic combustor configuration and then making 12 minor design changes to optimize the performance of this co&ustor. The results obtained from the experimental investiga- tion of the model 13 com3mstor at low pressures and high air-flow rates are summarized in the following paragraphs. The values quoted for sim- ulated flight performance refer to the model 13 combustor in a typic& 5.2 pressure ratio turbojet engine at a flight Mach number of 0.6 .

1. Although the combustor was developed for liquid fuel injectJon, heated liquid and vapor fuels injected through the conventional liquid- fuel injection system gave higher conkustion efficiencies at the more severe operating conditions. At conditions simulating cruise speed at ..

14 - NACA RM E52J09

4

80,000 feet altitude, the efficiencies with heated liquid and vapor fuels were 18 and 23 percent, respectively, above the efficiency with liquid fuel. A t cruise speed at- 70,000 fee t , the differences in efficiency between the fuels w e r e sl ight.

-

2. With a l l three fuel types investigated (liquid, heated liquid, N

and vapor) the combustion efficiency at severe operating conditions was higher than the efficiency of any of 14 turbojet combustors previously investigated. A t rated engine speed the combustion efficiency w a s above 97 percent with both liquid fuel and vapor fuel at a l t i tudes up t o 65,000 feet.

4 8

3. A t cruise speed at 56 ,OOO feet alt i tude, the conibustion eff i- ciency w&s slightly increased by increasing the air-flow rate per unit combustor frontal area t o a value 30 percent above that used in current engine design practice. A t 851 air-flow rate 69 percent above current practice, however, the efficiency was markedly decreased at- high fuel- air ra t ios .

4. A smaller, similar annular cortibustor which xae previously devel- oped to obtain high performance would not operate at many conditions at which the model 13. cmibustor produced high efficiency. The better per- formance of the model l3 combustor may therefore be due chief ly to Its size. A comparison of the combustion efficiency previously obtained w i t h 14 turbojet combustors a t :comparable operating conditions shows that fo r most of the combustors the efficiency increases with increase i n combustor size.

5. The total-pressure drop through the conibustor w&s approximately twice as great as the pressure drop of some production-model combustors.

6. The combustor-outlet temperatures followed the radial pattern generally desired i n turbojet qngjnes; that is, the temperatures were low at the extreme blade hub and blade t ip posi t ions and were a maximum a t about 85 percent of the blade height. The maximum circumferential vaxiation of individual thermocouple indications from the rean tempera- ture at any radial position was about +ZOO0 F.

7. The low-altitude performance of the combustor has not been inves- tigated; consequently, little is known regarding i ts durabili ty or carbon-deposition characteristics.

Lewis Flight Propulsion Laboratory Nat ional Advisory Committee fo r Aeronautics

Cleveland, Ohio -?eij'~.

.

NACA RM E52J09

i

15

APPENDIX A

EVOLUTION OF MODEL 13 COMBUSTOR

The design modiflcations consisted of two distinct types: (1) sec- ondary zone modifications, which prbar i ly affected the out le t - temperature prof i le and the combustor pressure drop; and (2) primary zone modifications, which primarily affected the conibustion efficiency. The greatest deficiency of the model 1 conbustor was I t s ou t le t - temperature profile; hence seconday zone mdiffcations w e r e made first.

Secondmy Zone Modifications

Figure 22.shows the combustor liner design fo r three conbustors (models I, 2, and 31, and figure 2 3 shows the radial outlet-temperature profiles of these combustors. Combustor mdels 1 t o 3 had approximately the same total. open area in the liner and hence the same pressure drop. Model1 produced temperatures which were tcm l o w near the inner w a l l (turbine blade h&) and too high near the center of the duct (fig. 23(a)).

Model 2 was evolved to improve the temperature profile of modell . This combustor had less open area on the inner wall of the liner and correspondingly greater open area on the outer wall, resulting in increased temperatures near the blade hub (fig. 23(b) ). In addition, the model 2 conibustor had a greater spacing between the air-admission s l o t s in the inner wall of the liner, permitting the cold air entering through these slots t o penetrate 86 indLvidual j e t s into the hot gases t o produce alternate hot and cold "corridors" in the conibustor. With the smaller s lo t spacing of the preceding combustor (modell), these ~ i r jets did not penetrate as individual Jets, but displaced the hot gases toward the center of the duct and coalesced to form a cold layer of gas near the turbine blade hub. These deductions are based on observations of the flame patterns Fn the combustor and the temperature patterns visi- ble on the side walls of the codmstor housing. The improved j e t pene- tration (obtained with the model 2 conibustor) se rvedto reduce the tem- perature in the center of the duct (fig. 23(b)).

The model 3 codnstor represented a s t i l l further step in the same direction; that is, less qpen area in the inner wall, correspondingly weater open area in the outer wall, and wider spacing between the slots in the inner wall. A further Fzsprovement in outlet-temperature profile resulted (fig. 23(c)).

An increase in the area of the secondary zone liner perforations was required to decrease the pressure drop of the model 3 combustor.

- The width of the secondary zone slots could not be increased without

16 . . - NACA RM E52J09

w l'educing the space between the .secondmy air jets; this would be expected t o impair the penetration of these jets as indicated by the experience w i t h combustor models 1 t o 3. The slots could not be extended upstream - without reducing the length available for conibustion. The requisite increase in area of the s l o t s was therefore obtained by changing *om N

10 slots (fig. 22) t o 5 s lots ( f ig . 6) ; this permitted a 30 percent increase i n the area of the secondary air s l o t s . i n model 13 over that of models 1 t o 3. The outlet-temperature profile was sl ight ly impaired by this area increase, but the accompanying decrease (approximately 25 per- cent) in pressure drop made the over-all pdormance of the model13 com- bustor more desirable.

4

3

After the change w-agmade from the 10-slot t o the 5-slot secondary zone configuration, three modifications were required to obtain the desired outlet-temperature profile by means of a correct balance between secondary s lo t area Fn-rtheauter w a l l and Fn the inner wall o f the liner (models ll t o 13).

Attempts to greatfly increase the area of the secondary alr s lo t s (models 5 and 6 ) resulted in circumferentially uneven temperature pro- f i les and severe local high-tenperahre reg iws . Wdels 5 and 6 also showed no marked lowering of the pressure drop, probably because a large par t of the pressure loss was then occurring i n the ann- flow pas- sages which supply air t o the downstream par t of the cambustor l iner.

I

Additional secondary zone modifications were t o add cooling louvers at the positions indicated

PrimaJry Zone Modifications

Combustor model 4 was similar t o model 3 but consisting of ten 3/16-inch holes per row located

made in models 7 and 8 - In figure 6.

."

had 3 rows of holes i n the inner and outer

w a l l s o e h e liner '(for a to ta l o f -s ix ty 3/16-Fnch holes in the one- q m t e r sector) approximately 7- inches f r o m t h e upstream end of the l iner. These small holes had no measurable effect on performance and were later replaced by a row of cooling louvers.

1 2

Combustor models 9 and 10 vazied the amount o f a i r admitted behind the fuel atomizer radiation shield (fig. 5). This air flows around the fue l atomizers and emerges i n t o the combustion zone through the holes provided i n the radiation shield for each f u e l atomizer. Model 9 had a row of 1/16-inch hoLes i n the .upstream end of the l iner w a l l s for admis- sion of t h i s afr behind the fuel atomizers; the f inal model 13 colnbustor had 1/8-inch diameter holes for t h i s purpose, and the mdel 10 combustor had alternate 1/8-inch and 5/32-inch holes for this purpose. Thus models 9 and 10 provided less.and more air, respectively, around the -

b

.. .

3B NACA RM E52J09 .- 17

P fuel atomizers than did model13. The conibustion efficiencies of models 9 and 10 were b e l o w that of model13, indicating that the air

further indicating that model 13 provided approximately the optimum quan- c admitted around the fuel nozzles has a marked effect on efficiency and

@3 4 tity of air in this location. 3

18 SACA RM E52J09

APPENDIX B t

SYM.EiOLS

The following symbols are used i n this report :

reference co&ustor crose-sectional =ea, sq f t

fue l -a i r ra t io , lb/lb

combustor-inlet total pressure, in. Hg absolute

total-pressure drop through combustor, in. Hg

fuel manifold pressure (above cambustor-inlet pressure), in. Hg

canbustor-inlet stattc pressure, lb/sq f t absolute

reference dynamic pressure, in. Hg conibustor-inlet total temperature, OR

mean combustor-outlet temperature, OR mean temperature rise through codmstor, 9

combustor reference velocity, ft/sec

air-fhw rate, U/sec

fuel-flow rate, ~3/~lr combustion efficiency, percent

combustor-inlet air density, lb/cu f t

coaustor-outlet air density, lb/cu f t

NACA RM E52J09 " 19

REFERENCES

1. Olson, Walter T., and Scbroeter, Thomas T.: Effect of Distribution of Basket-Eole Area on Simulated Altitude Performance of 25Z-Inch- Diameter Annular-Type Turbo jet Conibustor. NACA RM E8AO2, 1948.

1

2. Mark, Herman, and Zettle, Eugene V. : Effect of Air Distribution on Radial Temperature Distribution i n One-Sixth Sector of Annular Turbojet Conibustor. W A BM E9122, 1950.

3, Zettle, Eugene V. , and Mark, Herman: Simulated Altitude Performuice of Two Annular Cortibmtors with Continuous Axial Openings f o r Admis- sion of Primary A i r . IUCA RM E50E18a, 1950.

4. McCafferty, Richard J. : Vapor-Fuel-Distribution Effects on Combustor Performance of a Single Tubular Cortibustor. NACA RM E50J03, 1950.

5. McCafferty, Richard J. : Liquid-Fuel-Distribution and Fuel-State H f e c t s on Conibustion Performance of a Single Tubular Conibustor. NACA RM E51B21, 1951.

6. C h i l d s , J. Howard: Prelfminasy Correlation of Efficiency of A i r c r a f t Gas-Turbine Conibustors for Different Operating Conditions. NACA RM E50F15, 1950.

7. Harp, James L., Jr., and Vincent, Kenneth R. r Altitude Performance Investigation of S a l e - and Double-Annul= Turbojet-Engine Com- bustors with V a r i o u s Size Fuel Nozzles. W A R M E5U;14, 1952.

8. Turner, L. Richard, and Bogart, Donald: Constant-Pressure Conibustion C h a r t s Including Effects of.Diluent Addition. NACA Rep. 937, 1949. (Supersedes NACA TN's 1086 and 1655.)

9. Straight, David M., and Gernon, J. Dean: Photographic Studies of Preignition Environment and Flame In i t ia t ion in Turbojet-Engine Conibustors. NACA RM E52I l l .

10. Belles, Frank E., and Sfmon, Dorothy M.: Variation of the Pressure L imi t s of Flame Propagation with M e D i a m e t e r f o r Propane-Air Mixtures. NACA RM E51J09, 1951.

20 NACA FLM E52J09

TABIE I - FUEL ANALYSIS Fuel properties

A.S.T.M. dis t i l l a t ion D86-46, ?E' Initial boillng point Percentage evaporated

5 . '

lo 20 30 40 50 60 70 80 90

Final boiling point Residue, percent Loss, percent

Aromatics A.S.T.M. D-875-46T,

percent by volume Silica gel, percent

by volume Specific gravity Viscosity, centistokes

Reid vapor pressure,

Hydrogen-carbon r a t i o Net heat of combGtion

at looo F

lb/sq in.

MIL-F-5624A (Jp-4) ( M A fuel 52-53)

136

183 200 225 244 263 278 301 321 347 400 498 1.2 0.7

8.5

10.7

0.757 0.762

2.9

0.170 18,700

NACA RM E52J09 21

a, 0 cc cu' h l , Ip -

- I - 59.5 4-0 46.4 c.s Q.S 4 3 E S A 4.1 75.6 5.0

lo1.7 5.5 85.9 5.1

20.5 2.8

89.5 s.0 zs.8 2.8

55.0 3-8 40.3 3.5

. P . 2 3 3 BS.9 4-0

ZS.0 4.1 s2.0 4.S 57.6 4.J 45.2 4.5

41.4 - a6.8 - QI.0 - 51.0 - 2l.2 5 3 5 . 6 3.8 2l.5 5.3 27.0 4.0 32.4 L S 56.9 4.s

m.o 4 3

M

~

6.0

50

72J

725 7ss 7 s 730 m 728 725

729 7 s l

736 19) 722

722 735 72s 72s

n 2

754

750

10.6 Dayhr

78.84

8 O . U

80.61 80.87

50.0 m l b I 735

727

22 MACA RM E52J09

.ob w.0

116 11? .18 .19

.a1 w.06

15.0 . 15.1 16.4 15.5 L6.25 11.6 15.2 15.0 15.0

8.0 5.0

8.6

llJ U.0 1'0

51.1 283

9-7 10.1 9.4 D .E

10.7 ll.1 ls ,O

9 .l 9.7

10.1 1o.s

9 .6

9.a

8.1

128 "- I

2.694 1S1 .Ox10

5.004 w1.1 P.Wl m.9

3.- w1.0 5.072 wl.0 5.118 u 1 . 1 1.104 w1.0

lm.S*l 181.2 180.0

182.6

I""

"I-

5.074 189.2 1.812 m 1 . 4

. m 4 1 m . 5 3590 503.7

"

c

. . .

I

.. . . .

2709

Plenm

r Vent 7

1 4 1 .. .. . Altituly

N w

24 NACA I34 E52J09

.

(a) Inlet thrnOCOUpbS (iroZI-CanStant+), inlet total-pressure rakea, and stream-statio probe in plane at station 1.

Q Thermocouple 0 Total-pessure rake ~~Sta t i c -p res su re o r i f i ce [I Stream-static probe

(b) Outlet thermocouples (ohramel-alumel) in plane at station 2.

rigure 2. - Location6 of instnrmentstian.

4B

m 0 PC N

NACA RM E52J09 25

(b) Outlet thermocouple rake.

1 ( e ) Wedge stream-statlo probe.

'33-2646

60LZ

1 . . .

2709 . . . . . . . .

. .

I

NACA RM E52JO9

cn 0 ec N

1800

1600

CP

1200

1000

29

Hub 1 2 3 4 Tip Distance along turbine blade, in.

Q A 15 268 2.14 O B 8 268 1.14 A C 5 268 .71C

-

CN 0

I Fuel-air ratio

Figure 8. - Combustion efficiency of model 13 combustor w i t h liquid MLF-5624A eP.* JP-4 fuel at varloua pressures. 10.5-gallon-per-hour, 600 fuel atomizers. E

0 CD !G

1 I 60 LZ I '

. . . .. . . . . . . . .

2709

. . . . . . . . . . . . .

. . . . . . . . .. . - .. . . . . " .

!4

w N

I

, . . . . . . . . ... .

SOL2

- . . . . . . . . . . - . .

*

.. -

2709

34 NACA RM E52J09

100 "

90

80

Condl- PI - 'Ti , ' lid& ' 'Ion (In- Ed ("1 ((sec)(ss ,,)) 70 -

A D 15 2 68 2 . 7 0 9 E 15 2 68 3.62 - 1 - - - - Curve from fig. 8, condition A

60

50 I .006 .#a .010 : .012 .014 .016 .018 .a0

Fuel-air ratio

Figure 12. - Coaibustion efficiency of model 13 conibustor with liquid MILF-5624A grade JP-4 fuel at various air-flw rates. 10.5-gallon-per-hour, 60' fuel atomizers.

N 4 0 co

. . . . . . . .

- - - - Curve frcm f i n . 10, condition A I I -

.M36 .ma .010 .ol2 .014 .016 .010 Wel-air ratio

. . . . . . . . . . . . . . . .

2709

I

,020 .022

. . .

2 E

Figure 13. - Combustion efficiency of model 13 combustor wlth vapor (propane) fuel at varione sir- flow rates. Two fuel atomizer c a p c i t i m .

".

- . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . - . . . . . . .. . . - .

I

60 LZ

NACA RM E52J09 37

2000

1800

I

1600

.

800

600 Hub 1 2 3 4 Tip 5

Distance along turbine blade, In. (a) Liquid MIL-F-56248 grade JP-4 fuel.

=-97

Figure 15. - Combustor-outlet temperature prof i l e a .

38 NACA FM E5ZJ09

Figure 15- - Concluded. : Combustor-outlet temperstwe profiles.

. . . . . . .

I

. . . . . . . .

* I 2709

Fuel

50 u Vapor

.. . . .

#

Fucl-air ratio

Figure 16. - Conrparlaou at cornhuntion efficiency for three fuel states at variolls preseures.

. .

40 NACA €34 E52J09

.

100

90

-lJ

0

PI

8 3 5 80

E 9 4 0

€3 'O

3 Q V 0 B l iquid fuel .

al

T-L -P Condition . .

0 open synibols represent

60 A C 'Pailed symbols represent A D heated 1Lquid fuel . 9 E Solid synibols represent

vapor fuel.

P 50 "z97 80 120 160 . 200 240 280 328(10-6

Combustion parameter, Vr/piTI (ft, lb, seeJ % units)

(a) Temperature rise, 680' F . Figure 1 7 . - Correlation of combustion efficiency data of figures 8, 9.

10 12, and 13 w i t h combustion parameter V,/piTi-

1 - 1

t

6B NACA RM E52J09

(b) Temperature rise, 402' F.

Figure 17. - Continued. Correlation of cmibustion efficfency data of figures 10, 12, and 13 with conkustion parameter Vr/PPITI.

41

42 NACA RM E52J09

6

c

Combustion parameter, Vr/BITI (ft, lb, sec, OR units)

( c ) Temperature rise, 1180' F .

Figure 17. - Concluded. Corr&lation of conibustion efficiency data of figures 10, 12, and 13 with combustion parameter Vr/PiTi-

NACA RM E52J09

Cantration parameter V , / P ~ T ~ (ft, I%, sec, % units)

F u m e 18. - Comparison of curves of flgne 17 with similar curves for better combustors of reference E.

43

r c * 60 LZ

. . . . . . . . . . . . . .... -. . .

L

. . . . . . . . . ....

4

..

60 LZ *

R

I

46 0 NACA RM E52J09

1400

combustor previously investigated

and combustion effl- ciencies for madel 13

1300 0 Experinaental values

1200

1100

8 loo0

!2

3

.. % .A k 900

c, I 800

700

600

500

I 400.

50 100 150 200 . . 250 300 35Ok Conibuation parameter, V J P ~ T ~ ( f t , ~ b , ~ e c , % unite)

Figure 20. - Comparison of altitude perfonnance of model13 combustor with perfonnance of ~ i m i l ~ combustor of smaller s i ze .

I

r

*

0 0

.

10-

. . . . . . . . .

I

. . . .. .

a . 2709

. - . . . . . . . . . . . .

4

. - . . . . . . . . . . . . . . -. . . - . . . ... -.

/ \ I=" -7 /I \ i ! \ I

I \ i \ I

p. m

I

60LZ

I I

. . ........

* . . . . . . . . . . . . . . .

a iir

Inner wall

#

Outer wall

I

. . . . . . . . . . . . . . . . . . . . . , _. . . . ".. - - . . . . . . . . . . . . . .. . . . , . .

VI 0

( 0 ) Moas1 3 .

n #

t

NACA RM E52J09 - 51

1 2 3 4 m P 5 DiBtsnce along turbine blade, in.

(a) Model 1. -z97

Figure 23. - Combustor-outlet temperature profiles.

52

(b) Mcdel 2.

NACA RM E52J09

Figure 23. - Continued. Combustor-outlet temperature prof i l e a .

.

833

c

NACA RM E5WO9 53

Hub 2 3 4 Tip 5 Diatsnce along turbine blade, h.

(c) Model 3. v

NACA--ley - 1-15-53 - 4W

SECURITY INFORMATION

c

4